This repository contains the code and data used for the paper "Turnout and Amendment 4: Mobilizing Eligible Voters Close to Formerly Incarcerated Floridians"

Changing the "view" option to "Tree" is recommended in the Dataverse to better understand the file structure.

00_initialize.R and 01_geocode_releasees.R create outputs that are used throughout the analyses and should be run first

01_precinct_results, 02_treated_households, and 03_hills_rob all represent different analyses from the paper. These folders (and the scripts within them) should be run in order.

The "misc" folder includes the code for creating the LaTeX tables and for creating the parallelized cluster for the genetic matching.

The "helpers" folder includes functions from the author's personal package that are called in the analyses; where relevant, the scripts in the main analyses load these functions.

In an attempt to balance transparency with the need to protect the privacy of voters living with or near formerly incarcerated individuals, I have not included the raw data (though I do include the code used to get from the raw data to the regression data). I am instead including the following datasets:

1. "temp/precint_level_reg_data.rds". This is the dataset on which the precinct-level regressions (and marginal effects plots) are run in script 01_precinct_level_results/04_precinct_bg_regs.R. The precinct id has been dropped.

2. "temp/bg_level_reg_data.rds". This is the dataset on which the precinct-level regressions (and marginal effects plots) are run in script 01_precinct_level_results/04_precinct_bg_regs.R. The GEOID has been dropped.

3. "temp/pre_reg_av.rdata". This is the result of the post-matching exercise on voters living with formerly incarcerated individuals. These include anonymized voter ID numbers. These data are the input for 02_treated_households/09_run_regressions.R.

4. "temp/bg_level_hills_prereg.rds" and "temp/results_demos_hills_pre_reg.rds" provide the data for the neighborhood-level Hillsborough County analyses (for use in 03_hills_rob/02_precinct.R); "temp/pre_reg_hills.rdata" provides the individual-level (anonymized) data for Hillsborough County regressions (for use in 03_hills_rob/treated_households/07_matching_reg.R).